International Journal of Engineering and Technical Research (IJETR)

ISSN: 2321-0869 (O) 2454-4698 (P), Volume-7, Issue-2, February 2017

Experimental Analysis of a Swirl Burner for

Propulsion Applications: Influence of Thermal and
Fluid Dynamic Field on Pollutant Emissions
Giulio Solero

Abstract— Non-premixed swirl burners are widely used in
technical appliances (such as propulsion, gas turbines, boilers)
by virtue of high flame stability, mainly due to the generation of
a central recirculation region characterized by efficient mixing
between the reactants and rapid homogenisation of the
combustible mixture. Swirl motion imparted to the air flow
presents a strong influence upon combustion features (i.e.: flame
morphology, thermal and fluid dynamic field) and pollutant
emissions. In spite of the wide use, many aspects of swirling
reacting flows still have to be thoroughly investigated:
experimental measurements are difficult owing to high
turbulence levels and possible onset of instability phenomena
and, consequently, numerical simulation of this flow typology is
far to provide reliable results, especially at high Reynolds
number of the reacting flow. This paper presents the
experimental results obtained comparing different natural gas
injection typologies in a swirl burner. Particularly, both co-axial
and radial (i.e.: transverse) injection, with respect to the
rotating air stream, have been characterised through different
techniques: particle image velocimetry and laser Doppler
anemometry for flow field analysis, temperature measurements
by thin thermocouple and pollutant emissions measurement at
the exhaust. The results put into evidence that, although the
global mixing process is mainly governed by the swirling air
stream, in the region close to the reactants efflux the fuel
injection procedure plays an important role for flame
stabilization and development in the primary mixing zone of the
device. Moreover, the general behaviour of the two different
injectors (mainly as for pollutant emissions) seems to reflect the
generation of two different flame typologies: a partially
premixed one for the radial injector and a purely diffusive flame
for the axial one.
Index Terms— combustion, swirl burners, pollutant
emissions
I. INTRODUCTION
Non-premixed swirling flows are widely used in industrial
combustion systems, particularly propulsion, gas turbines,
boilers and furnaces, for safety and stability reasons [1]. Swirl
motion of the air flow increases flame stability and has strong
influence on the combustion efficiency and on the pollutant
emissions [2]. The basic principle of the swirl flow is that
above a certain swirl level (S >0.6), there is the generation of
a recirculation bubble in the proximity of the fuel jet outlet,
the so called CTRZ (Central Toroidal Recirculation Zone)
[3]. The combustion process is strongly influenced by the
dimension and shape of the recirculation zone, because the
combustion products recirculate backwards and supply
energy for the ignition of the incoming fuel-air stream [4].
Moreover, the recirculating regime provides efficient mixing
between the reactants and a rapid homogenisation of the
combustible mixture [5]: the CTRZ behaves as a well mixed
source for heat and free radicals transfer, allowing self
sustainment, stabilization and propagation of the flame.
In spite of the wide use, swirling reacting flows are far from
being fully understood, both under the point of view of
experimental measurements and numerical simulation, owing
to high flow complexity and possible onset (at high swirl
intensity) of instability phenomena, such as the PVC
(Precessing Vortex Core).
This paper deals with the experimental characterization of a
natural gas swirl combustor, analysing by different techniques
(flame visualization, PIV and LDA, temperature and pollutant
emissions measurement) the influence of fuel injection
typology (coaxial or transverse with respect to the swirling air
stream) and swirl intensity upon the flame behaviour
(morphology, thermal and flow field, environmental impact).
II. EXPERIMENTAL SET-UP
Fig. 1 reports a schematic view of the investigated burner,
which can be considered as a prototype for propulsion and gas
turbine applications (for more details, see [6-7]).
As it can be seen, the burner is equipped with an
axial+tangential swirl generator: it is a configuration widely
used in typical engineering appliances (swirl intensity can be
easily varied through the axial-tangential split ratio). A
cylindrical quartz combustion chamber (internal
diameter=192 mm) has been used for flame confinement,
making possible flame visualization and measurements by
optical techniques. A natural draught hood provides the
exhaust and sampling of the burned gases.
As previously outlined, the burner can be equipped
alternatively with two fuel injector typologies: the co-axial
injector (with respect to the air stream) presents a 8 mm
circular single nozzle; the radial injector provides fuel
admission transversal to the air stream and has been designed
with eight circular holes so as to reproduce (with respect to
the axial one) similar Reynolds number at the efflux.

Giulio Solero, Department of Energy - Politecnico di Milano via

Fig. 1: the analised swirl burner.
Lambruschini, 4 - 20156 Milano – Italy

13 www.erpublication.org
 Experimental Analysis of a Swirl Burner for Propulsion Applications: Influence of Thermal and Fluid Dynamic
Field on Pollutant Emissions
Tab. 1 reports the nominal operating conditions used for the
experimental measurements described in this paper.

Air flow rate [g/s] 8.8

Reynolds number of air jet 20700
Natural gas flow rate [g/s] 0.35
Reynolds number of natural gas jet 5600
Input thermal power [kW] 17
Air swirl number S 0.82
Fuel/Air Momentum ratio MR 0.92
Fuel/Air Equivalence ratio  0.69
Tab.1: nominal operating conditions of the burner.
Fig. 2a: typical image of the flame for axial injector.
Owing to the high complexity of the flow inside the
combustion chamber (due to the high swirl intensity imparted
to the air stream), the experimental characterization has been
carried out through different techniques.
Flow field measurements in reacting conditions have been
performed both by laser Doppler anemometry and particle
image velocimetry [8]. In this case, a double seeding has been
used to have the complete characterization of the flow field.
Particularly:
 Silicon oil droplets (mean diameter=1
dispersed in the fuel flow, to reproduce the central jet
penetration (in the case of axial injector) and obtain velocity
measurements referred to the “cold” natural gas jet
interacting with the recirculating central bubble;
 Alumina particles (mean diameter=5
added in the corner region at the base of the combustion
chamber, reproducing mainly the recirculating flow of Fig. 2b: typical image of the flame for radial injector.
already burned gases.
Mean temperature was measured using a Pt/Pt-13% Rh bare
wire thermocouple with 0.3 mm diameter bead. The amplified As for the macroscopic flow field, the difference between the
signals were sampled at a 500 Hz sampling frequency and the two injectors is visible in Fig. 3a, b, which reports the 2-D
mean value was based on 5000 instantaneous data. A flow field measured by PIV averaging 200 double-exposed
correction was made for the radiation error, following [9] and images (for the radial injector, the investigated plane includes
using the measured velocity values for the evaluation of two injection nozzles). For both injectors, it is clearly
convective heat transfer coefficient. Finally, burned gases noticeable the generation of the recirculation bubble due to
have been sampled for analysis of the pollutant emissions swirl effect imparted to the air stream. However, the use of the
(chemiluminescence for NOx and infrared analysis for CO). radial injector, obviously, avoids the possible interaction of
the central jet with the formation of the recirculating region,
which is generated just downstream the efflux, very close to
III. RESULTS AND DISCUSSION the burner head, contributing to flame stability and reactants
mixing (with already burned gases too) enhancement with
3.1 – Flame morphology and flow field respect to axial injection. In fact, fuel jets seem to be soon
entrained in the transverse air stream, inducing rapid mixing.
Fig. 2 a, b reports the image of the flame for the different For the axial injector, the interaction between the central fuel
injection typologies (nominal operating conditions reported jet and the recirculating central region has been deepened
in Tab. 1). It can be observed the typical calyx shaped flame, through LDA. Velocity measurements for the axial
due to swirl, and, for the axial injector, the formation of a component along the burner axis and close to the region of
central luminous region connected to fuel penetration inside interaction gave rise to bi-modal distributions (see Fig. 4),
the recirculating bubble, generating a fuel rich zone and connected to the simultaneous presence of the positive (fuel
giving rise to soot formation. This phenomenon, anyway jet) and negative (recirculating gases) velocity regime. This
sporadic for axial injector, is always absent for the radial one. puts into evidence the possibility of sporadic penetration of
The higher stability and compactness of the flame in the case the fuel jet inside the bubble, a phenomenon originating the
of radial injection is proved by the results obtained by CH* luminous zone visible in Fig. 2a. In Fig. 4 the progressive
emission spectroscopy from the flame front, not reported here appearance of the central recirculation region and its
[10], putting into evidence that the reaction zone (identified interaction with the central fuel jet is clearly visible.
by the peak of CH* emission intensity) for radial injector is
closer and more concentrated at the burner head, with an
initial steeper gradient.

14 www.erpublication.org
 International Journal of Engineering and Technical Research (IJETR)
ISSN: 2321-0869 (O) 2454-4698 (P), Volume-7, Issue-2, February 2017

1500
1400
1300
1200
1100
1000
900

T [°C]
800
700
600
500 H 3 mm
H 18 mm
400
H 33 mm
300 H 63 mm
200 H 93 mm
100 H 123 mm
0
0 1 2 3 4 5
R/Reff
Fig. 3a: Flow field for axial injector.
Fig. 5: radial temperature semi-profiles measured for
axial injector.

1300
1200
1100
1000
900
Tem peratura [C]

800
700
600
500
H 3 mm
400
H 18 mm
300 H 33 mm
200 H 63 mm
H 93 mm
100 H 123 mm
0
Fig. 3b: Flow field for radial injector. 0 1 2 3 4 5
R/R eff

Fig. 6: radial temperature semi-profiles measured for

radial injector.

Close to the efflux, the axial injector presents a central

relatively “cold” region which can be attributed to the natural
gas jet, separated from the parallel “cold” air stream by a
relatively hot zone that can be attributed to incipient
MR=0.92, =0.69, h/d=3.25. MR=0.92, =0.69, h/d=3.5. formation of the recirculation region and development of
combustion reactions. At the contrary, for the radial injector
the central region is characterised by high temperature levels
due to recirculation of hot already burned gases till the burner
head (see Fig. 3b); air stream outflowing is clearly visible
with low temperature levels (similar for the two injection
typologies). At the periphery (i.e., R/Reff>1) in both cases
there is a region characterised by a quasi uniform temperature
value (about 800-1000 °C) which is connected to the
formation of a corner recirculation zone visible also in Figgs.
MR=0.92, =0.69, h/d=3.75. MR=0.92, =0.69, h/d=4. 3a, b. This corner reverse flow, already observed in a similar
burner [11], is induced by the air stream radial expansion and
Fig. 4: velocity histograms from LDA measurements (axial the wall confinement: the high temperatures measured in this
injector) at progressive increasing distance from the efflux. zone indicate the presence of a large amount of already
3.2 – Thermal field burned gases that are entrained by the reactant flow.

The different behaviour of the two injectors, especially close
to the reactants efflux, is evident also in the temperature
measurements. Figgs. 5, 6 report the comparison of radial
temperature semi-profiles, at progressive increasing distance
H from the efflux, for axial and radial injector respectively.
Figgs. 7, 8 present the super-imposition of mean temperature
and mean axial velocity profiles at a distance H=3 mm from
the efflux, confirming the correspondence of thermal and flow
field, previously described.

15 www.erpublication.org
 Experimental Analysis of a Swirl Burner for Propulsion Applications: Influence of Thermal and Fluid Dynamic
Field on Pollutant Emissions
However, the radial injector gives rise to more uniform
900 30

800
profiles and differences up to 150 °C are present probably
connected to a distinct development of combustion reactions
25

700
20
and local heat release (see Figgs. 2a, b and 3).

Axial Vel. [m/s]

600
Temperature [°C]

15
500

400
10
3.3 – Pollutant emissions
300
5 Finally, pollutant emissions (CO and NOx) at the exhaust
200
Temperature
0 have been measured for the two injectors in different
100
Axial velocity -5 operating conditions, that is varying equivalence ratio, and for
0 -10 two values of air swirl number: the nominal one (0.82) and a
lower value (0.65). Swirl number variation is possible
0 1 2 3 4 5
R/R eff

modifying the axial-tangential split ratio in the swirl

Fig. 7: super-imposition of mean temperature and axial generator. Variation of the equivalence ratio has been
velocity profile at H=3 mm for the axial injector. obtained changing fuel flow rate and, consequently, input
1000 30
thermal power and momentum ratio, but maintaining constant
the air flow rate and, as a consequence, the Reynolds number
25
800 and the macroscopic fluid dynamic of the flow. Results are
reported in Figgs. 11, 12, 13 and 14. As it can be seen, the
20
Axial Vel. [m/s]
Temperature [°C]

600
15
graphs report the blow-off limit for the lean flame and are also
10
useful to define the possible operability range of the burner.
Blow-off limit in lean conditions is mainly dictated by swirl
400
5

200 Temperature
0
intensity rather than injection procedure: in this case, the
Axial velocity -5 lower swirl number allows the extension of blow-off limit
0
0 1 2 3 4 5
-10 towards leaner conditions, to the detriment of CO emissions
R/R eff which become very high. Probably, the higher value of swirl
number can enhance flame local stretching inducing
Fig. 8: super-imposition of mean temperature and axial instability phenomena (such as the PVC, Precessing Vortex
velocity profile at H=3 mm for the radial injector. Core) clearly observed in isothermal conditions. The radial
injector presents lower NOx emissions (up to 50% in lean
Figgs. 9, 10 report the comparison of temperature-axial condition) with respect to the axial one and its behaviour
velocity trend along the burner axis, for both injectors. under the point of view of pollutant emissions is strictly
15,0
dependent from equivalence ratio rather than swirl number
1200 12,5 (although enhancement of swirl intensity gives rise to lower
1000
10,0
emission levels). In fact, for the radial injector, a steep
increase of NOx emissions has been pointed out approaching
7,5
Axial Vel. [m/s]

5,0

stoichiometric flames. At the same time, a relevant increase of

Temperature [°C]

800
2,5

600
0,0

-2,5
CO formation is noticeable in lean conditions. At the
400
-5,0 contrary, especially for high swirl number, the axial injector
-7,5 seems quite insensitive to equivalence ratio and a slight
decrease in NOx emissions can be revealed close to
200 -10,0
Temperature
Axial Velocity -12,5

0 -15,0 stoichiometric flame, associated with increase of CO emission

0 10 20 30 40 50 60 70 80 90 100 110 120 130
H [mm] (probably connected to strong penetration of central fuel jet
with subsequent possible mixing deficiency).
Fig. 9: temperature and axial velocity behaviour along The general behaviour of the two different injectors as for
the burner axis, for axial injector. pollutant emissions seems to reflect the generation of two
different flame typologies: a partially premixed one for the
radial injector and a purely diffusive flame for the axial one.
15,0
12 00 12,5
Radial
10,0 SWIRL 0.65 Axial
10 00 7,5
110
Axial Vel. [m/s]

5,0 100
Temperature [°C]

8 00
2,5
90
0,0 blow-off
6 00
-2,5
mg NO2/Nm3 3% O2

80
-5,0
4 00
70
-7,5

2 00 -10,0 60
T e m p e ra tu re
A xia l V e lo city -12,5
50
0 -15,0
0 10 20 30 40 50 60 70 80 90 100 11 0 1 20 130
40
H [m m ] blow-off
30

Fig. 10: temperature and axial velocity behaviour along 20

0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 Equiv. Ratio

the burner axis, for radial injector. 12

0.32
15
0.47
20
0.9
25.2
1.41
Input Power[kW]
MR

At increasing distance from the efflux the temperature profiles

become more uniform (full development of combustion Fig. 11: comparison of NOx emissions at swirl number=0.65.
reactions): the profile trend is similar for the two injectors.

16 www.erpublication.org
 International Journal of Engineering and Technical Research (IJETR)
ISSN: 2321-0869 (O) 2454-4698 (P), Volume-7, Issue-2, February 2017
fuel penetration and formation of a sooting luminous region, a
phenomenon obviously absent in the case of radial injection.
Radial
blow-off
Axial
SWIRL 0.65
12000 The difference between the two injection procedures is clearly
10000
noticeable also as for pollutant emissions at the exhaust. In
fact, the radial injector presents lower emission levels with
respect to axial one and seems to give rise to a flame quite
mg/Nm3 CO 3% O2

8000

6000
similar to a partially premixed one, with positive effects on
development of new burners characterised by low
4000 environmental impact.
blow-off
2000
Moreover, the obtained results can constitute a representative
0
0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 Equiv. Ratio
data set for validation of numerical codes and turbulence
12
0.32
15
0.47
20
0.9
25.2
1.41
Input Power [kW]
MR models in the field of reacting turbulent flows.

Fig. 12: comparison of CO emissions at swirl number=0.65.

Radial REFERENCES
110 SWIRL 0.82 Axial
[1] A.K. Gupta, D.G. Lilley, N. Syred: “Swirl flows”, Abacus Press,
100
Tunbridge Wells, 1984
90 [2] T.C. Claypole, N. Syred: “The effect of swirl burner aerodynamics on
NOx formation”, 18th Symposium (International) on Combustion, The
80
Combustion Institute, 1981, p. 81
mg NO2/Nm3 3% O2

70 [3] R. Hillemans, B. Lenze, W. Leuckel: “Flame stabilization and turbulent

exchange in strongly swirling natural gas flames”, 21 st Symposium
60 blow-off
(International) on Combustion, The Combustion Institute, 1986, p. 1445
50 [4] V. Tangirala, J.F. Driscoll: “Temperatures within non-premixed flames:
40
effects of rapid mixing due to swirl”, Combustion Science and
Technology, 60, (1988), 143
30 blow-off [5] R.H. Chen, J.F. Driscoll: “The role of recirculation vortex in improving
20 fuel-air mixing within swirling flames”, 22nd Symposium (International)
0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 Equiv. Ratio
Input Power[kW]
on Combustion, The Combustion Institute, 1988, p. 531
12 15 20 25.2
0.32 0.47 0.9 1.41 MR [6] G. Solero, A. Coghe, G. Scribano: “Influence of natural gas injection
procedure in a swirl burner”, Joint Meeting of the Greek and Italian
Sections of The Combustion Institute, Corfu, June 17-19 2004
Fig. 13: comparison of NOx emissions at swirl number=0.82. [7] Olivani A., Solero G., Cozzi F., Coghe A., (2007): “Near field flow
structure of isothermal swirling flows and reacting non-premixed
Radial
Axial
swirling flames”, Experimental Thermal and Fluid Science, vol. 31,
2500
SWIRL 0.82 427-436
[8] Araneo L., Coghe A., Cozzi F., Olivani A., Solero G., (2008): “Natural
gas burners for domestic and industrial appliances: application of the
2000
particle image velocimetry (PIV) technique”, in: Schroder A., Willert
C.E., Particle Image Velocimetry, Topics in Appl. Physics, pp. 245-257,
Heidelberg Springer Verlag
mg/Nm3 CO 3% O2

1500
[9] T.V. Morgan: “Thermal behaviour of electrical conductors”, John Wiley
and Sons, New York, 1991
1000 [10] G. Solero, A. Olivani, F. Cozzi, A. Coghe (2005): “Experimental
analysis of fuel injection procedure in a natural gas swirling flame”,
European Combustion Symposium, 2005
500
blow-off
[11] A. Coghe, G. Solero, G. Scribano (2004): “Recirculation phenomena in
a natural gas swirl combustor”, Experimental Thermal and Fluid
blow-off
Science, vol. 28, (2004), 709-714
0
0,3 0,4 0,5 0,6 0,7 0,8 0,9 1,0 Equiv. Ratio
12 15 20 25.2 Input Power [kW]
0.32 0.47 0.9 1.41 MR

Fig. 14: comparison of CO emissions at swirl number=0.82.

IV. MAIN CONCLUSIONS

The experimental analysis through different techniques of a

natural gas burner varying the gas injection procedure (axial
or transverse with respect to air stream) put into evidence how
this procedure plays an important role in flame stabilization
and development close to the reactants efflux, being the
global mixing process governed by swirl effect imparted to
the air. Particularly, it has been deepened the knowledge
about the possible interaction (for axial injection) between the
central fuel jet and the recirculating bubble, which can induce

17 www.erpublication.org